Project 3: Dissecting and evolving the mating module of budding yeast, S. cerevisiae (Murray) (#21-26) We used experiment and theory to ask how yeast cells polarize and pick mating partners. Cell polarization is essential for development and plays an important role in the biology of many human pathogens. Yeast cells polarize up pheromone gradients but also polarize in spatially uniform pheromone fields. We ruled out two models for polarization, the use of a historical mark, and models based on lateral inhibition. Our results support a new class of model, global integration, which uses actin filaments and positive feedback to integrate signaling throughout the cell: pheromone signaling induces actin polymerization, and secretory vesicles move along actin filaments to deliver more signaling components. Cells can only detect pheromone gradients over a narrow concentration range. They use the regulated secretion of Bar1, a protease that degrades pheromone, to keep the pheromone concentration at the cell surface in this range. In collaboration with Naama Barkai, we have shown that although exogenous protease substantially improves the mating of Bar1-deficient cells, it does not allow them to discriminate between identically attractive partners, a process that requires cell surfacebound Bar1. We used the mating module to investigate the evolutionary pliability of modules. Much experimental evolution has focused on metabolic traits like improved growth on limiting nutrient sources, but we chose to evolve or engineer traits that depended on a signaling module: 1) We developed a laboratory model of sympatric speciation and used it to evolve a five-fold mating preference30. We developed general methods to map and identify these mutations, with the goal of revealing the genetic and phenotypic basis of speciation31 2) We tried to evolve cells where DNA damage would activate a reporter of pheromone signaling, but instead produced strains in which reporter activity varies stochastically within a cell lineage. This trait depends on 4 mutations of which we have mapped two and are close to identifying the other two. 3) We engineered the graded pheromone response into a hysteretic switch by putting modified signaling proteins under the control of a pheromone-inducible promoter. This work shows that expression of a single gene can determine whether a pathway responds reversibly to an environmental signal, acts as a bistable switch, or is constitutively activated32. We began work on the fundamental parameters that control evolution. Attempting to evolve the mating module prompted us to ask fundamental questions about evolution. These included searching for theory that would relate the rate of evolution to fundamental parameters (population size, mutation rate, interaction between mutations, etc.) and devising ways to measure these parameters. In collaboration with M. Desai (a Physics graduate student) and D. Fisher, we developed and tested a model that predicted the rate of evolution33, and we investigated the conditions under which mutators accelerated evolution34. We also made detailed measurements of the per-base pair mutation rate35 and showed that the local average of this rate changes across the genome in concert with the timing of DNA replication, with the latest replicating sequences having the highest mutation rate. Further experiments show that early replicating sequences are less likely to be repaired by error-prone DNA polymerases than late replicating sequences.